Elamur Research⌕
☰
home / papers / 2026 / Identifying Bias in Machine-generated Text Detection

Identifying Bias in Machine-generated Text Detection

Paper · 2026 · Kevin Stowe, Svetlana Afanaseva, Rodolfo C. Raimundo et al.

Summary

Detectors that flag text as machine-generated are increasingly used in schools and content moderation. When they wrongly label human writing as AI-produced, the harm may fall unevenly on vulnerable groups. This paper checks English detectors for that kind of unfairness. The authors merged and cleaned three US student essay corpora (PERSUADE V 2.0, ASAP V 2.0, ELLIPSE) for grades 6–12, which include writer demographics. They then ran 16 public detectors, both zero-shot and trained, on these human-written essays. Following the framework of Dayanik et al. (2022), they used regression to test four attributes (gender, race/ethnicity, ELL status, economic status) while controlling for confounds. They also examined demographic subgroups.

No single bias pattern holds across all detectors. Still, several detectors flag disadvantaged writers more often. ELL essays tend to be misclassified, especially those by non-White students. Essays by economically disadvantaged students are, unexpectedly, flagged less often. Three expert human annotators scored near chance but showed no significant bias. The authors conclude that each detector needs its own fairness audit before deployment.

Key points

  • The 16 detectors include zero-shot systems (Ghostbuster, Glimpse, Binoculars, two Fast-DetectGPT variants, Zippy) and trained ones (four BiScope variants, three DeTeCtive variants, Desklib, e5-lora, RADAR).
  • The DeTeCtive OUTFOX variant was left out because its data overlaps with the evaluation essays.
  • Among race subgroups, most models flag A/PI essays more often and B/AA essays less often. The AI/AN group (n = 208) was too small to show significant effects.
  • Human accuracy was 0.449–0.526, and no attribute effect was significant at p < .01.
  • Limitations include essays as the only domain, binary gender labels, predefined racial categories, and only three annotators.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksSafety Research · 543 works

Cite

Cite as Kevin Stowe, Svetlana Afanaseva, Rodolfo C. Raimundo, Yitao Sun and Kailash Patil. 2026. Identifying Bias in Machine-generated Text Detection. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.109

@inproceedings{stowe2026identifying,
  title = {{Identifying Bias in Machine-generated Text Detection}},
  author = {Kevin Stowe, Svetlana Afanaseva, Rodolfo C. Raimundo, Yitao Sun and Kailash Patil},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.109},
  url = {https://doi.org/10.18653/v1/2026.acl-long.109}
}

References

45 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Firoj Alam, Preslav Nakov, Nizar Habash, Iryna Gurevych, Shammur Chowdhury, Artem Shelmanov, Yuxia Wang, Ekaterina Artemova, Mucahid Kutlu, and George Mikros, editors. 2025. Proceedings of the 1stWorkshop on GenAI Content Detection (GenAIDe-tect). International Conference on Computational Linguistic… find
  2. Anthropic. 2024. Claude 3.5 sonnet. https://www.anthropic.com/claude. ↗
  3. Guangsheng Bao, Yanbin Zhao, Juncai He, and Yue Zhang. 2025. Glimpse: Enabling white-box methods to use proprietary models for zero-shot llm-generated text detection. Preprint, arXiv:2412.11506. — [arXiv:2412.11506](https://arxiv.org/abs/2412.11506) find
  4. Guangsheng Bao, Yanbin Zhao, Zhiyang Teng, Linyi Yang, and Yue Zhang. 2024. Fast-detectgpt: Efficient zero-shot detection of machine-generated text via conditional probability curvature. Preprint, arXiv:2310.05130.
  5. Marion Bartl, Abhishek Mandal, Susan Leavy, and Suzanne Little. 2025. Gender bias in natural language processing and computer vision: A comparative survey. ACM Comput. Surv., 57(6). find
  6. Laura Biester. 2025. Sports and women’s sports: Gender bias in text generation with olympic data. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 2: Short Papers), pages 195–205, Albuqu… find
  7. Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. 2020. Language (technology) is power: A critical survey of “bias” in NLP. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5454– 5476, Online. Association for Computational Linguistics.
  8. Elizabeth Clark, Tal August, Sofia Serrano, Nikita Haduong, Suchin Gururangan, and Noah A. Smith. 2021. All that’s ‘human’ is not gold: Evaluating human evaluation of generated text. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International… find
  9. S.A. Crossley, Y. Tian, P. Baffour, A. Franklin, M. Benner, and U. Boser. 2024. A large-scale corpus for assessing written argumentation: Persuade 2.0. Assessing Writing, 61:100865. find
  10. Scott Crossley. 2024. The english language learner insight, proficiency and skills evaluation (ellipse) corpus. International Journal of Learner Corpus Research, 9(2). find
  11. Scott A. Crossley, Perpetual Baffour, L. Burleigh, and Jules King. 2025. A large-scale corpus for assessing source-based writing quality: Asap 2.0. Assessing Writing, 65:100954. find
  12. Paula Czarnowska, Yogarshi Vyas, and Kashif Shah. 2021. Quantifying social biases in nlp: A generalization and empirical comparison of extrinsic fairness metrics. Transactions of the Association for Computational Linguistics, 9:1249–1267.
  13. Erenay Dayanik, Ngoc Thang Vu, and Sebastian Padó. 2022. Bias identification and attribution in NLP models with regression and effect sizes. Northern European Journal of Language Technology, 8. find
  14. Desklib. 2025. ai-text-detector: Ai-generated text detection model. Accessed: 2025-07-10. find
  15. Sunipa Dev, Masoud Monajatipoor, Anaelia Ovalle, Arjun Subramonian, Jeff Phillips, and Kai-Wei Chang. 2021. Harms of gender exclusivity and challenges in non-binary representation in language technologies. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pag…
  16. Liam Dugan, Alyssa Hwang, Filip Trhlík, Andrew Zhu, Josh Magnus Ludan, Hainiu Xu, Daphne Ippolito, and Chris Callison-Burch. 2024. RAID: A shared benchmark for robust evaluation of machine-generated text detectors. In Proceedings of the 62nd Annual Meeting of the Association for Computational Lingui…
  17. Liam Dugan, Daphne Ippolito, Arun Kirubarajan, Sherry Shi, and Chris Callison-Burch. 2023. Real or fake text? investigating human ability to detect boundaries between human-written and machine-generated text. In Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-… find
  18. Economic Policy Institute. 2022. Racial and ethnic disparities in the united states: An interactive chartbook. find
  19. Updated November 2024. find
  20. Kawin Ethayarajh and Dan Jurafsky. 2022. The authenticity gap in human evaluation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 6056–6070, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find
  21. Anjalie Field, Su Lin Blodgett, Zeerak Waseem, and Yulia Tsvetkov. 2021. A survey of race, racism, and anti-racism in NLP. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1:… find
  22. Hanxi Guo, Siyuan Cheng, Xiaolong Jin, Zhuo Zhang, Kaiyuan Zhang, Guanhong Tao, Guangyu Shen, and Xiangyu Zhang. 2024a. Biscope: Ai-generated text detection by checking memorization of preceding tokens. In Advances in Neural Information Processing Systems (NeurIPS), volume 37, pages 104065– 104090. find
  23. Xun Guo, Shan Zhang, Yongxin He, Ting Zhang, Wanquan Feng, Haibin Huang, and Chongyang Ma. 2024b. Detective: Detecting ai-generated text via multi-level contrastive learning. Preprint, arXiv:2410.20964. — [arXiv:2410.20964](https://arxiv.org/abs/2410.20964) find
  24. Hans W. A. Hanley and Zakir Durumeric. 2024. Machine-made media: Monitoring the mobilization of machine-generated articles on misinformation and mainstream news websites. Preprint, arXiv:2305.09820. — [arXiv:2305.09820](https://arxiv.org/abs/2305.09820) find
  25. Alex Hanna, Remi Denton, Andrew Smart, and Jamila Smith-Loud. 2020. Towards a critical race methodology in algorithmic fairness. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, FAT* ’20, page 501–512, New York, NY, USA. Association for Computing Machinery. find
  26. Abhimanyu Hans, Avi Schwarzschild, Valeriia Cherepanova, Hamid Kazemi, Aniruddha Saha, Micah Goldblum, Jonas Geiping, and Tom Goldstein. 2024. Spotting llms with binoculars: zero-shot detection of machine-generated text. In Proceedings of the 41st International Conference on Machine Learning, ICML’2… find
  27. Xiaomeng Hu, Pin-Yu Chen, and Tsung-Yi Ho. 2023. Radar: Robust ai-text detection via adversarial learning. In Advances in Neural Information Processing Systems, volume 36, pages 15077–15095. Curran Associates, Inc. real: Deepfake moderation mistakes and identity-based bias. In Proceedings of the 202… find
  28. Srinivasan Iyer, Xi Victoria Lin, Ramakanth Pasunuru, Todor Mihaylov, Daniel Simig, Ping Yu, Kurt Shuster, Tianlu Wang, Qing Liu, Punit Singh Koura, Xian Li, Brian O’Horo, Gabriel Pereyra, Jeff Wang, Christopher Dewan, Asli Celikyilmaz, Luke Zettlemoyer, and Ves Stoyanov. 2023. Opt-iml: Scaling lang…
  29. Yikang Pan, Liangming Pan, Wenhu Chen, Preslav Nakov, Min-Yen Kan, and William Wang. 2023. On the risk of misinformation pollution with large language models. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 1389–1403, Singapore. Association for Computational Linguisti…
  30. Yan Ju, Shu Hu, Shan Jia, George H. Chen, and Siwei Lyu. 2024. Improving fairness in deepfake detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 4655–4665. find
  31. Minseok Jung. 2025. Responsible computational text generation: Ai content classification and policy framework. Master’s thesis, Massachusetts Institute of Technology, Cambridge, MA, February. find
  32. Vinaya Sree Katamneni, Aakash Varma Nadimpalli, and Ajita Rattani. 2024. Demographic fairness and accountability of audio- and video-based unimodal and bi-modal deepfake detectors. Face Recognition Across the Imaging Spectrum, pages 205–231. find
  33. Ryuto Koike, Masahiro Kaneko, and Naoaki Okazaki. 2024. Outfox: Llm-generated essay detection through in-context learning with adversarially generated examples. Proceedings of the AAAI Conference on Artificial Intelligence, 38(19):21258–21266. find
  34. Jooyoung Lee, Xiaochen Zhu, Georgi Karadzhov, Tom Stafford, Andreas Vlachos, and Dongwon Lee. 2025. Collaborative evaluation of deepfake text with deliberation-enhancing dialogue systems. Preprint, arXiv:2503.04945. — [arXiv:2503.04945](https://arxiv.org/abs/2503.04945) find
  35. Flor Miriam Plaza-del Arco, Amanda Cercas Curry, Alba Curry, Gavin Abercrombie, and Dirk Hovy. 2024. Angry men, sad women: Large language models reflect gendered stereotypes in emotion attribution. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1:… find
  36. Felipe Romero-Moreno. 2025. Deepfake detection in generative ai: A legal framework proposal to protect human rights. Computer Law & Security Review, 58:106162. find
  37. Beatrice Savoldi, Marco Gaido, Luisa Bentivogli, Matteo Negri, and Marco Turchi. 2021. Gender bias in machine translation. Transactions of the Association for Computational Linguistics, 9:845–874.
  38. Karolina Stanczak and Isabelle Augenstein. 2021. A survey on gender bias in natural language processing. Preprint, arXiv:2112.14168. — [arXiv:2112.14168](https://arxiv.org/abs/2112.14168) find
  39. Thinkst. 2023. Meet zippy: A fast ai/llm text detector. Accessed: 12/03/2025. find
  40. Weixin Liang, Mert Yuksekgonul, Yining Mao, Eric Wu, and James Zou. 2023. Gpt detectors are biased against non-native english writers. Patterns, 4(7):100779. Accessed: 2025-03-12.
  41. Vivek Verma, Eve Fleisig, Nicholas Tomlin, and Dan Klein. 2024. Ghostbuster: Detecting text ghostwritten by large language models. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)…
  42. Decheng Liu, Zongqi Wang, Chunlei Peng, Nannan Wang, Ruimin Hu, and Xinbo Gao. 2025. Thinking racial bias in fair forgery detection: Models, datasets and evaluations. Proceedings of the AAAI Conference on Artificial Intelligence, 39(5):5379–5387. find
  43. Junchao Wu, Shu Yang, Runzhe Zhan, Yulin Yuan, Lidia Sam Chao, and Derek Fai Wong. 2025. A survey on LLM-generated text detection: Necessity, methods, and future directions. Computational Linguistics, 51(1):275–338.
  44. Jesse G Meyer, Ryan J Urbanowicz, Patrick CN Martin, Karen O’Connor, Ruowang Li, Pei-Chen Peng, Tiffani J Bright, Nicholas Tatonetti, Kyoung Jae Won, Graciela Gonzalez-Hernandez, and Jason H Moore. 2023. Chatgpt and large language models in academia: opportunities and challenges. BioData Mining, 16(… find
  45. A Jaron Mink, Miranda Wei, Collins W. Munyendo, Kurt Hugenberg, Tadayoshi Kohno, Elissa M. Redmiles, and Gang Wang. 2024. It’s trying too hard to look Data Statistics for the human-written dataset that was curated for our experiments are shown in Table 7. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/stowe-2026-identifying-bias-machine-generated-text/